<p>With the development of edge computing, a growing volume of service requests is directed towards edge servers for processing. To ensure the successful execution of service requests, corresponding services need to be deployed on edge servers. However, deploying services on these servers presents challenges due to resource constraints, server heterogeneity, and device mobility. Although there are many deployment options, there are few studies on service deployment issues in shipyard environments. To solve the above problems, this paper considers the importance of edge servers in the network and the impact when deploying the same services. We propose an improved genetic algorithm based on center priority and clustering in a fixed equipment environment. Additionally, considering the mobility of terminal devices, the paper first predicts the future service requests of the devices using Long Short-Term Memory (LSTM) and then proposes a dual clustering algorithm to formulate a service deployment plan. Additionally, this paper positions the nearest service table on the edge server to minimize the query time for routing service requests to other servers. Compared with the benchmark algorithms, CPIGA reduces the delay by 8.45%<InlineEquation ID="IEq1"> <InlineMediaObject> <ImageObject Color="BlackWhite" FileRef="10586_2025_5150_Article_IEq1.gif" Format="GIF" Height="13" Rendition="HTML" Resolution="72" Type="Linedraw" Width="19" /> </InlineMediaObject> <EquationSource Format="TEX">$-$</EquationSource> <EquationSource Format="MATHML"><math> <mo>-</mo> </math></EquationSource> </InlineEquation>12.2% and reduces the cost by 36%<InlineEquation ID="IEq2"> <InlineMediaObject> <ImageObject Color="BlackWhite" FileRef="10586_2025_5150_Article_IEq1.gif" Format="GIF" Height="13" Rendition="HTML" Resolution="72" Type="Linedraw" Width="19" /> </InlineMediaObject> <EquationSource Format="TEX">$-$</EquationSource> <EquationSource Format="MATHML"><math> <mo>-</mo> </math></EquationSource> </InlineEquation>38.5%. LSTM-DCA reduces the delay by 15.5%–25% and improves the success rate of service one request by 13%–26%. The experimental results show that the algorithm proposed in this paper outperforms the benchmark algorithm in terms of latency and cost.</p>

An improved heuristic service deployment algorithm based on multi-parameter priority in edge computing

期刊论文
发表日期：

2025年7月31日

查看全文

Zhenyu Yin,
Zhiying Bi,
Junyao Xue,
Lei Wu

摘要

With the development of edge computing, a growing volume of service requests is directed towards edge servers for processing. To ensure the successful execution of service requests, corresponding services need to be deployed on edge servers. However, deploying services on these servers presents challenges due to resource constraints, server heterogeneity, and device mobility. Although there are many deployment options, there are few studies on service deployment issues in shipyard environments. To solve the above problems, this paper considers the importance of edge servers in the network and the impact when deploying the same services. We propose an improved genetic algorithm based on center priority and clustering in a fixed equipment environment. Additionally, considering the mobility of terminal devices, the paper first predicts the future service requests of the devices using Long Short-Term Memory (LSTM) and then proposes a dual clustering algorithm to formulate a service deployment plan. Additionally, this paper positions the nearest service table on the edge server to minimize the query time for routing service requests to other servers. Compared with the benchmark algorithms, CPIGA reduces the delay by 8.45% $-$ $-$ 12.2% and reduces the cost by 36% $-$ $-$ 38.5%. LSTM-DCA reduces the delay by 15.5%–25% and improves the success rate of service one request by 13%–26%. The experimental results show that the algorithm proposed in this paper outperforms the benchmark algorithm in terms of latency and cost.